Short answer: Test consent as a sequence, not a static yes/no value. The order in which defaults are applied, tags evaluate them, and updates arrive can change measurement behavior. Test with a clean browser state and repeat after each relevant change.
1. Establish the intended policy first
Before testing, define the organization's required default states by geography, consent-management platform and measurement purpose. The correct legal and business defaults vary. Do not label a site non-compliant solely because an event was observed: Google's advanced Consent Mode can send limited, cookieless measurements under denied storage states.
2. Open an isolated session
Use a private window or a clean test profile. Clear relevant consent choices, set the intended geography, and start Google Tag Assistant on a page you can safely inspect. Capture the initial consent state before interacting with the banner.
3. Check all relevant dimensions
Verify the values for analytics_storage, ad_storage, ad_user_data, and ad_personalization. Confirm that the default was set before the tags depending on it evaluated consent. Google recommends its consent APIs and Consent Initialization for GTM-based setups. Google's official Consent Mode debugging guide describes the test sequence and common failures.
4. Repeat with accept, reject and custom choices
- Capture initial state and what traffic, if any, was observed.
- Accept each relevant category and confirm the expected update precedes subsequent tag behavior.
- Reset and reject, checking what the chosen Consent Mode implementation is expected to send.
- Test navigation between pages and preservation of user choices.
- Repeat for the geographical policies or subdomains the business actually uses.
5. Avoid a common false positive
A denied analytics storage signal does not always mean zero network traffic in advanced mode. The meaningful question is whether the implementation behaves according to its stated consent policy and Google's supported behavior. Google's Consent Mode overview explains how basic and advanced implementations differ.
6. Make regression tests reproducible
Save expected defaults, update transitions, browser conditions, destination rules, and relevant screenshots in a change-controlled checklist. Retest after CMP, GTM, runtime, or marketing-platform changes. A browser tool can help observe behavior, but it is not a legal compliance certificate.
Related: What the browser can verify.